Prettify rescuetime provider, interact with export module

my/rescuetime.py

@@ -0,0 +1,71 @@
+import logging
+from pathlib import Path
+from datetime import datetime, timedelta
+from typing import NamedTuple, Dict, List, Set, Optional
+from functools import lru_cache
+
+from .common import get_files
+
+# TODO get rid of it
+from kython import group_by_cmp # type: ignore
+
+from my_configuration import paths
+
+
+def get_logger():
+    return logging.getLogger("my.rescuetime")
+
+
+def _get_exports() -> List[Path]:
+    from my_configuration import paths
+    return get_files(paths.rescuetime.export_path, '*.json')
+
+
+import my_configuration.repos.rescuexport.model as rescuexport
+Model = rescuexport.Model
+
+
+# TODO cache?
+def get_model(last=0) -> Model:
+    return Model(_get_exports()[-last:])
+
+
+def get_groups(gap=timedelta(hours=3)):
+    model = get_model()
+    it = model.iter_entries()
+    lit = list(it) # TODO get rid of it...
+    return group_by_cmp(lit, lambda a, b: (b.dt - a.dt) <= gap, dist=1)
+
+
+def print_groups():
+    for gr in get_groups():
+         print(f"{gr[0].dt}--{gr[-1].dt}")
+    # TODO merged db?
+    # TODO ok, it summarises my sleep intervals pretty well. I guess should adjust it for the fact I don't sleep during the day, and it would be ok!
+
+
+def check_backed_up():
+    model = get_model(last=1)
+    last = list(model.iter_entries())[-1]
+    latest_dt = last.dt
+
+    assert (datetime.now() - latest_dt) < timedelta(days=1)
+    # TODO move this to backup checker??
+
+
+def fill_influxdb():
+    from influxdb import InfluxDBClient # type: ignore
+    client = InfluxDBClient()
+    # client.delete_series(database='lastfm', measurement='phone')
+    db = 'test'
+    client.drop_database(db)
+    client.create_database(db)
+    model = get_model()
+    jsons = [{
+        "measurement": 'phone',
+        "tags": {},
+        "time": str(e.dt),
+        "fields": {"name": e.activity},
+    } for e in model.iter_entries()]
+    client.write_points(jsons, database=db) # TODO??
+

my/rescuetime/__init__.py

@@ -1,103 +0,0 @@
-import logging
-from pathlib import Path
-import json
-from datetime import datetime, timedelta
-from typing import NamedTuple, Dict, List, Set, Optional
-from functools import lru_cache
-
-from ..common import get_files
-
-
-from kython import JSONType, fget, group_by_cmp
-
-
-def get_logger():
-    return logging.getLogger("rescuetime-provider")
-
-_PATH = Path("/L/backups/rescuetime")
-
-def try_load(fp: Path):
-    logger = get_logger()
-    try:
-        return json.loads(fp.read_text())
-    except Exception as e:
-        if 'Expecting value' in str(e):
-            logger.warning(f"Corrupted: {fp}")
-        else:
-            raise e
-    return None
-
-
-_DT_FMT = "%Y-%m-%dT%H:%M:%S"
-
-class Entry(NamedTuple):
-    # TODO ugh, appears to be local time...
-    dt: datetime
-    duration_s: int
-    activity: str
-
-    @staticmethod
-    def from_row(row: List):
-        COL_DT = 0
-        COL_DUR = 1
-        COL_ACTIVITY = 3
-        dt_s     = row[COL_DT]
-        dur      = row[COL_DUR]
-        activity = row[COL_ACTIVITY]
-        # TODO utc??
-        dt = datetime.strptime(dt_s, _DT_FMT)
-        return Entry(dt=dt, duration_s=dur, activity=activity)
-
-
-# TODO hmm ok, these are going to need to be properly merged?
-@lru_cache(1)
-def get_rescuetime(latest: Optional[int]=None):
-    if latest is None:
-        latest = 0
-
-    entries: Set[Entry] = set()
-
-    # pylint: disable=invalid-unary-operand-type
-    for fp in list(sorted(_PATH.glob('*.json')))[-latest:]:
-        j = try_load(fp)
-        if j is None:
-            continue
-
-        cols = j['row_headers']
-        seen = 0
-        total = 0
-        for row in j['rows']:
-            e = Entry.from_row(row)
-            total += 1
-            if e in entries:
-                seen += 1
-            else:
-                entries.add(e)
-        print(f"{fp}: {seen}/{total}")
-        # import ipdb; ipdb.set_trace()
-        # print(len(j))
-    res = sorted(entries, key=fget(Entry.dt))
-    return res
-
-
-def get_groups(gap=timedelta(hours=3)):
-    data = get_rescuetime()
-    return group_by_cmp(data, lambda a, b: (b.dt - a.dt) <= gap, dist=1)
-
-
-
-def fill_influxdb():
-    from influxdb import InfluxDBClient # type: ignore
-    client = InfluxDBClient()
-    # client.delete_series(database='lastfm', measurement='phone')
-    db = 'test'
-    client.drop_database(db)
-    client.create_database(db)
-    jsons = [{
-        "measurement": 'phone',
-        "tags": {},
-        "time": str(e.dt),
-        "fields": {"name": e.activity},
-    } for e in get_rescuetime()]
-    client.write_points(jsons, database=db) # TODO??
-

my/rescuetime/__main__.py

@@ -1,19 +0,0 @@
-from kython.klogging import setup_logzero
-
-from . import get_logger, get_groups, get_rescuetime, fill_influxdb
-
-def main():
-    logger = get_logger()
-    setup_logzero(logger)
-
-    # for gr in get_groups():
-    #     print(f"{gr[0].dt}--{gr[-1].dt}")
-    # for e in get_rescuetime(latest=2):
-    #     print(e)
-    fill_influxdb()
-
-    # TODO merged db?
-    # TODO ok, it summarises my sleep intervals pretty well. I guess should adjust it for the fact I don't sleep during the day, and it would be ok!
-
-if __name__ == '__main__':
-    main()

my/rescuetime/check

@@ -1,15 +0,0 @@
-#!/usr/bin/env python3
-from datetime import datetime, timedelta
-
-from rescuetime import get_rescuetime
-
-
-def main():
-    rs = get_rescuetime(latest=1)
-    latest_dt = rs[-1].dt
-
-    assert (datetime.now() - latest_dt) < timedelta(days=1)
-
-
-if __name__ == '__main__':
-    main()